Deepfake Fraud
The use of AI-generated synthetic video or audio that realistically imitates a real person, deployed to deceive victims into authorising payments, sharing secrets, or trusting fake content.
Also known as: AI impersonation fraud, deepfake scam, synthetic media fraud, video cloning fraud
Last reviewed: 1 June 2026
Deepfake fraud applies AI image and audio synthesis to create convincing video or voice impersonations of real individuals. In financial fraud, the most impactful use cases to date have been live video calls where an executive's face is replaced in real time to authorise a fraudulent wire transfer, audio-only calls using voice-cloned executives to instruct employees or family members to transfer funds, and fake video 'endorsements' from celebrities used to promote investment or cryptocurrency scams.
The technology has become accessible and affordable: free or low-cost tools can generate convincing voice clones from minutes of audio and face-swap video in near real time. This removes the historical constraint that limited scalable fraud to text and audio, making video verification no longer a reliable authentication method in isolation.
Countermeasures include establishing code words with family or colleagues to verify unexpected urgent requests, using out-of-band verification for financial transactions, and deploying deepfake-detection tools at organisational level. Awareness that deepfakes exist and are used in fraud is itself a significant defensive step.
Examples
- A finance employee joins a video call where a convincing real-time deepfake of the CFO instructs them to wire funds urgently for a confidential acquisition.
- A social-media ad features a deepfaked celebrity endorsing a cryptocurrency platform; clicking the ad leads to a scam investment site.